Posttraumatic cubitus varus deformity is the most common late complication of a supracondylar fracture or a physeal fracture of the distal aspect of the humerus. Humeral osteotomy is used to correct this deformity and to avoid later complications, such as ulnar nerve palsy1,2, posterolateral rotatory instability3,4, and secondary distal humeral fracture5,6. It is difficult to precisely correct a cubitus varus deformity to mirror the normal side because the deformity includes hyperextension and internal rotation. A variety of osteotomies has been proposed to correct the complex deformity, including lateral closing-wedge, medial opening-wedge7, dome8-11, pentalateral12,13, three-dimensional14,15, and step-cut osteotomies16-18. A surgical approach to correct internal rotation malalignment was first reported in 195919, followed by several later reports on the need to correct internal rotation14,15,20. The lateral closing-wedge osteotomy, which has had satisfactory results, and a step-cut osteotomy, described recently and also reported to yield good results18, correct the varus deformity in only a single coronal plane, and not three-dimensionally. To date, no direct comparison has been made between three-dimensional and simple coronal plane osteotomies.
We use distal humeral osteotomies for correction only in the coronal plane or in a three-dimensional manner. This study is a retrospective review of the cases of eighty-six patients divided into two groups according to the type of osteotomy done to address the question of whether the treatment of posttraumatic cubitus varus deformity should also include correction of hyperextension and internal rotation deformity.
This article is based on and reanalyzes a study first reported in Japanese in 200721.
Patients
Between 1983 and 2007, eighty-six patients with posttraumatic varus deformity were treated at our hospitals. When the difference between the affected side and the contralateral, intact side was =20°, a corrective osteotomy was done. There were fifty-nine male and twenty-seven female patients. The average age of the patients at the time of the osteotomy was 11.1 years (range, three to thirty-one years). The average follow-up period was 35.4 months (range, twenty-four to 140 months). The onset of the deformity was related to a distal humeral supracondylar fracture in sixty-two patients; an elbow injury, the details of which were unknown, in nineteen patients; and a physeal fracture of the distal aspect of the humerus in five patients. The institutional review board at the Keio University School of Medicine approved the retrospective observational study.
Surgical Approaches
Of the eighty-six patients, forty underwent a simple lateral closing-wedge osteotomy; thirty-four, a three-dimensional osteotomy; eight, a modified step-cut osteotomy; and four, a rotational dome osteotomy. In situ decompression or subcutaneous anterior transposition of the ulnar nerve in patients with tardy ulnar nerve palsy and reconstruction of the lateral collateral ligament in patients with tardy posterolateral rotatory instability were also done.
Three-Dimensional Osteotomy
The three-dimensional osteotomy was done as described by Usui et al.14. An incision was made on the lateral side of the distal end of the upper arm, exposing the distal part of the humerus. A lateral wedge osteotomy of the humerus was then performed. Next, the distal part of the arm was rotated along its axis to correct the internal rotation and was bent so as to enable the fingertips of the affected side to touch the shoulder on the same side, in order to correct the hyperextension (Fig. 1, A).
Rotational Dome Osteotomy
For the rotational dome osteotomy8, the distal end of the humerus was exposed through a posterior approach, splitting the triceps brachii muscle. The curved osteotomy was about 1.5 cm proximal to the tip of the olecranon with the apex of the dome proximal. After the osteotomy was completed, the distal fragment was manipulated in the coronal, sagittal, and axial planes to correct the varus, hyperextension, and internal rotation malunion (Fig. 1, B).
Simple Lateral Closing-Wedge Osteotomy
A simple lateral closing-wedge osteotomy was carried out on the lateral side of the distal end of the humerus similarly to the three-dimensional osteotomy. This osteotomy was done to correct only the varus deformity (Fig. 1, C).
Modified Step-cut Osteotomy
A modified step-cut osteotomy as described by Kim et al.17 and Yun et al.18 used a posterior approach, with the triceps brachii muscle split to expose the posterior aspect of the distal end of the humerus. The initial osteotomy was done 0.5 to 1 cm superior to the olecranon fossa and perpendicular to the lateral rim of the distal end of the humerus. A triangular template was placed over the humerus and the proximal cut of the humerus was then done as determined by the template. The lateral edge of the distal fragment was moved to the apex of the proximal osteotomy site, and the degree of correction increased as the apex was moved medially (Fig. 1, D).
More than two Kirschner wires were used as the method of fixation in all of the osteotomies, except for six patients who underwent a three-dimensional osteotomy and were more than twenty years old. In these patients, plate fixation was used to provide more rigid support.
Assessment of the Outcome After Correction
We designated patients undergoing the former two types of three-dimensional osteotomy (thirty-four patients who had the three-dimensional osteotomy and four who had the rotational dome osteotomy) as Group I and those undergoing the simple coronal plane osteotomies (forty patients who had the simple lateral closing-wedge osteotomy and eight who had the modified step-cut osteotomy) as Group II, and compared the outcomes between the groups.
The evaluation included assessment of the carrying angle from anteroposterior radiographs and measurement of the passive range of elbow motion with use of a goniometer before surgery and at the time of the final follow-up. We recorded complications such as infection, recurrence of cubitus varus deformity, nerve palsies, instability of the elbow, and refracture. Loss of correction was determined on the basis of the radiographic change in the carrying angle from immediately after the correction to the time of the final follow-up.
Loss-of-correction values were subjected to multiple linear regression analyses with model terms for surgery group, age, calendar time, sex, duration of follow-up, and interaction of surgery group with the duration of follow-up. The duration of follow-up was used for this analysis in order to compare group differences at a common value for follow-up duration.
To evaluate the remodeling capacity of the bone for the recovery of flexion, we assessed Group II, which did have the hyperextension surgically corrected, before surgery and at the time of the final follow-up, comparing the outcome between patients who were less than ten years old and those who were more than ten years old.
Statistical Analysis
All values are given as the mean and the standard error of the mean. The Fisher exact test was used to compare categorical data between the groups, and the Student t test was used to compare continuous data between the groups. The range of motion before surgery and at the final follow-up evaluation was compared with use of the paired t test. Differences were considered significant when the p value was <0.05.
Source of Funding
We did not receive any outside funding or grants in support of the research for or preparation of this work.
Preoperative Findings
The preoperative and postoperative clinical parameters are shown in Table I.
Group I comprised thirty-eight patients. The mean age at the time of the osteotomy was 12.3 years (range, three to thirty-one years), and twenty-seven patients (71%) were male. The onset of the deformity was related to a distal humeral supracondylar fracture in thirty patients and to a transphyseal fracture in two. Five patients developed a tardy ulnar nerve palsy, and four developed tardy posterolateral rotatory instability.
Group II comprised forty-eight patients. The mean age at the time of the osteotomy was 10.1 years (range, three to twenty-eight years), and thirty-two patients (67%) were male. The deformity was related to a distal humeral supracondylar fracture in thirty-two patients and to a transphyseal fracture in three. Seven patients developed late complications, including a tardy ulnar nerve palsy in four patients and posterolateral rotatory instability in three.
Postoperative Findings
No significant difference was detected between the groups in regard to the carrying angle or the range of motion, either before surgery or at the time of the final follow-up (Fig. 2, A, B, and C). However, Group I showed more significant loss of correction (p = 0.018). There was a loss of 3.6° from immediately after surgery to the time of the final follow-up in Group I compared with a loss of 0.7° in Group II (Table I) (Fig. 2, C).
Multiple regression analysis indicated that the difference in loss of correction between the groups remained even after adjusting for age, calendar time, sex, and follow-up duration. None of these additional factors were associated with loss of correction: age (p = 0.47), calendar time (p = 0.95), sex (p = 0.44), follow-up duration (p = 0.98), and interaction of surgery group with the follow-up duration (p = 0.85). Analogous multiple regression analyses for flexion, extension, and carrying angles predicted by surgery group, follow-up duration, and their interaction indicate that accounting for follow-up variation makes essentially no difference in the predicted values of these angles during the follow-up period. Likewise, multiple regression of flexion and extension angles on age category, follow-up duration, and their interaction make little or no difference in the predicted angles at the time of follow-up.
Complications After the Corrective Osteotomy
No patient had a postoperative infection in either group. The cubitus varus deformity recurred in three patients in Group I and in one patient in Group II. A transient iatrogenic radial nerve palsy was caused by lateral traction of the triceps brachii muscle belly in one patient in each group. Lateral instability and subluxation were caused in one patient in Group I by unknown factors. No patient had a late complication such as ulnar nerve palsy, posterolateral rotatory instability, or refracture (Table I).
Remodeling Capacity for the Recovery of Elbow Flexion
There were significant differences in the total arc of motion, including flexion, before surgery and at the final follow-up evaluation, in patients less than ten years old (but not those who were more than ten years old) in Group II. The extension was —1° and flexion was 133° at the time of the final follow-up in patients who were less than ten years of age (Table II) (Fig. 3).
Necessity for the Correction of Hyperextension
Gadgil et al.22 reported that patients older than ten years of age who were treated with straight-arm traction after distal humeral supracondylar fracture had some terminal restriction of flexion, indicating limited remodeling capacity in this age group. We also noted no significant remodeling in patients over ten years of age without the surgical correction of flexion. Thus, in patients under ten years of age, surgical correction of hyperextension is not needed when correcting the varus deformity.
Necessity for the Correction of Internal Rotation
Group-I patients who underwent three-dimensional correction showed more significant loss of correction from the immediate postoperative position to the position at the final evaluation. This difference may have occurred because the rotation reduced the bone contact area in the supracondylar region, making the osteotomy more unstable and difficult to pin23,24. Excessive derotation may lead to the formation of an anterior bulge with restriction of flexion23. If the rotational deformity is left untreated, complete correction of the posttraumatic cubitus varus deformity cannot be achieved14,15,25. However, the rates of recurrence and late complications were not significant in our study, even when the correction of internal rotation was ignored. None of the patients complained of difficulties with activities of daily living because of the loss of rotation. It appears that residual internal rotation deformity is well tolerated, although a limitation of this study is that patient satisfaction with the results of the osteotomy was not assessed. Another limitation of this study is that our conclusion cannot be applied to patients with physeal damage that involves a growth disturbance.
Cosmesis and Disability in Activities of Daily Living
Although posttraumatic cubitus varus presents as a deformity, this is attributable to the varus angulation per se and not the internal rotation malalignment. Rotational deformity is not the primary disability in residual cubitus varus because it is easily compensated for by rotation of the shoulder joint9.
Tardy Ulnar Nerve Palsy
Mitsunari et al. showed that an internal rotation deformity contributes to the onset of tardy ulnar nerve palsy26. With correction of the varus deformity, anterior compression of the nerve by the medial head of the triceps brachii is relieved. We believe it is sufficient to perform anterior transposition of the nerve at the time of the osteotomy in patients with preoperative symptoms related to the ulnar nerve27. In our patients, none had an ulnar nerve palsy preoperatively, and we did not encounter any patients with tardy ulnar nerve palsy, even among those in whom the internal rotation was not corrected.
Tardy Posterolateral Rotatory Instability
O'Driscoll et al.4 reported that displacement of the triceps brachii, which causes the instability, results from the varus deformity rather than from the internal rotation. In the present series with simple coronal plane correction, no patient developed tardy posterolateral rotatory instability, although it is not possible to know whether residual internal rotation alone can cause instability.
Secondary Fracture
Lateral condylar fracture5,28, lateral epicondylar avulsion fracture4, distal epiphyseal fracture5, and distal epicondylar fracture6 of the humerus following posttraumatic cubitus varus have been reported previously. Davids et al.28 suggested that varus alignment could increase both the distraction and shear forces across the lateral condyle of the distal end of the humerus generated by a routine fall on an outstretched upper arm. In a physeal fracture, the healed injury leaves the metaphysis thickened, which protects the area from further injury, but the physis becomes vulnerable. Internal rotation adds little risk for second fractures5.
In conclusion, in the surgical treatment of posttraumatic cubitus varus deformity, correction of internal rotation malalignment may not be needed as it is difficult to maintain the corrected carrying angle because of the small area of osseous contact. It is possible to treat tardy ulnar nerve palsy with anterior transposition and to treat tardy posterolateral rotatory instability with ligament reconstruction, even if the internal rotation is not corrected. Although there is a need to correct hyperextension in patients over the age of ten years, it is sufficient to correct varus deformity in only the coronal plane in those less than ten years of age to allow more precise and stable correction.